bm-20250208-3.14.0a4+-29f8a67-NOGIL

Results

• fork: python/29f8a67ae00081a36fdc
• version: 3.14.0a4+
• config: NOGIL
• commit hash: 29f8a67
• commit date: 2025-02-08T23:35:28+00:00
• commit merge base: c1f352bf0813803bb795b796c16040a5cd4115f2
• ref: 29f8a67ae00081a36fdc

linux aarch64 (arminc)

• GitHub Action run
• cpu model: missing
• platform: Linux-5.15.0-101-generic-aarch64-with-glibc2.35
• raw results

vs. 3.10.4

• Geometric mean: 1.105x faster (HPT: reliability of 99.91%, 1.01x faster at 99th %ile)
• Memory usage: 1.58x
• missing benchmarks: aiohttp, chameleon, dask, flaskblogging, gunicorn, mypy2, tornado_http
• new benchmarks: async_tree_cpu_io_mixed_tg, async_tree_io_tg, async_tree_memoization_tg, async_tree_none_tg, bpe_tokeniser, connected_components, k_core, many_optionals, shortest_path, sphinx, subparsers, unpack_sequence
• 📄table
• 📈time plot

vs. 3.12.0

• Geometric mean: 1.121x slower (HPT: reliability of 100.00%, 1.10x slower at 99th %ile)
• Memory usage: 1.26x
• missing benchmarks: aiohttp, chameleon, dask, gunicorn, mypy2, tornado_http
• new benchmarks: bpe_tokeniser, connected_components, k_core, many_optionals, shortest_path, sphinx, subparsers, unpack_sequence
• 📄table
• 📈time plot

vs. 3.13.0

• Geometric mean: 1.124x slower (HPT: reliability of 100.00%, 1.09x slower at 99th %ile)
• Memory usage: 1.24x
• missing benchmarks: chameleon, djangocms, gevent_hub, gunicorn, tornado_http
• new benchmarks: asyncio_tcp, asyncio_tcp_ssl, dulwich_log, pickle, pickle_dict, pickle_list, unpack_sequence, unpickle, unpickle_list
• 📄table
• 📈time plot

vs. base

• Geometric mean: 1.160x slower (HPT: reliability of 100.00%, 1.16x slower at 99th %ile)
• Memory usage: 1.21x
• 🧠memory plot
• 📄table
• 📈time plot

linux x86_64 (linux)

• GitHub Action run
• cpu model: Intel(R) Xeon(R) W-2255 CPU @ 3.70GHz
• platform: Linux-5.4.0-205-generic-x86_64-with-glibc2.31
• raw results

vs. 3.10.4

• Geometric mean: 1.267x faster (HPT: reliability of 100.00%, 1.13x faster at 99th %ile)
• Memory usage: 1.52x
• missing benchmarks: aiohttp, chameleon, dask, djangocms, flaskblogging, gunicorn, mypy2, tornado_http
• new benchmarks: async_tree_cpu_io_mixed_tg, async_tree_io_tg, async_tree_memoization_tg, async_tree_none_tg, bpe_tokeniser, connected_components, k_core, many_optionals, shortest_path, sphinx, subparsers
• 📄table
• 📈time plot

vs. 3.12.0

• Geometric mean: 1.016x slower (HPT: reliability of 99.84%, 1.00x slower at 99th %ile)
• Memory usage: 1.32x
• missing benchmarks: aiohttp, chameleon, dask, gunicorn, mypy2, tornado_http
• new benchmarks: bpe_tokeniser, connected_components, genshi_text, genshi_xml, html5lib, k_core, many_optionals, pylint, shortest_path, sphinx, subparsers, thrift
• 📄table
• 📈time plot

vs. 3.13.0

• Geometric mean: 1.079x slower (HPT: reliability of 100.00%, 1.05x slower at 99th %ile)
• Memory usage: 1.23x
• missing benchmarks: chameleon, djangocms, gevent_hub, gunicorn, tornado_http
• new benchmarks: asyncio_tcp, asyncio_tcp_ssl, pickle, pickle_dict, pickle_list, unpack_sequence, unpickle, unpickle_list
• 📄table
• 📈time plot

vs. base

• Geometric mean: 1.127x slower (HPT: reliability of 100.00%, 1.11x slower at 99th %ile)
• Memory usage: 1.20x
• 🧠memory plot
• 📄table
• 📈time plot

linux x86_64 (pythonperf2)

• GitHub Action run
• cpu model: 12th Gen Intel(R) Core(TM) i9-12900
• platform: Linux-5.15.0-125-generic-x86_64-with-glibc2.35
• raw results

vs. 3.10.4

• Geometric mean: 1.205x faster (HPT: reliability of 100.00%, 1.09x faster at 99th %ile)
• Memory usage: 1.54x
• missing benchmarks: aiohttp, chameleon, dask, flaskblogging, gunicorn, mypy2, tornado_http
• new benchmarks: async_tree_cpu_io_mixed_tg, async_tree_io_tg, async_tree_memoization_tg, async_tree_none_tg, bpe_tokeniser, connected_components, k_core, many_optionals, shortest_path, sphinx, subparsers
• 📄table
• 📈time plot

vs. 3.12.0

• Geometric mean: 1.057x slower (HPT: reliability of 99.94%, 1.03x slower at 99th %ile)
• Memory usage: 1.27x
• missing benchmarks: aiohttp, chameleon, dask, gunicorn, mypy2, tornado_http
• new benchmarks: bpe_tokeniser, connected_components, genshi_text, genshi_xml, html5lib, k_core, many_optionals, pylint, shortest_path, sphinx, subparsers, thrift
• 📄table
• 📈time plot

vs. 3.13.0

• Geometric mean: 1.049x slower (HPT: reliability of 99.86%, 1.01x slower at 99th %ile)
• Memory usage: 1.24x
• missing benchmarks: chameleon, djangocms, gevent_hub, gunicorn, tornado_http
• new benchmarks: asyncio_tcp, asyncio_tcp_ssl, pickle, pickle_dict, pickle_list, unpack_sequence, unpickle, unpickle_list
• 📄table
• 📈time plot

vs. base

• Geometric mean: 1.104x slower (HPT: reliability of 100.00%, 1.10x slower at 99th %ile)
• Memory usage: 1.21x
• 🧠memory plot
• 📄table
• 📈time plot

darwin arm64 (darwin)

• GitHub Action run
• cpu model: missing
• platform: macOS-15.3-arm64-arm-64bit-Mach-O
• raw results

vs. 3.10.4

• Geometric mean: 1.168x faster (HPT: reliability of 99.98%, 1.02x faster at 99th %ile)
• Memory usage: 1.29x
• missing benchmarks: chameleon, dask, djangocms, gevent_hub, tornado_http
• new benchmarks: async_tree_cpu_io_mixed_tg, async_tree_eager_cpu_io_mixed_tg, async_tree_eager_io_tg, async_tree_eager_memoization_tg, async_tree_eager_tg, async_tree_io_tg, async_tree_memoization_tg, async_tree_none_tg, asyncio_tcp, asyncio_tcp_ssl, pickle, pickle_dict, pickle_list, unpack_sequence, unpickle, unpickle_list
• 📄table
• 📈time plot

vs. 3.12.0

• Geometric mean: 1.073x slower (HPT: reliability of 99.99%, 1.03x slower at 99th %ile)
• Memory usage: 1.22x
• missing benchmarks: chameleon, dask, djangocms, gevent_hub, gunicorn, tornado_http
• new benchmarks: asyncio_tcp, asyncio_tcp_ssl, pickle, pickle_dict, pickle_list, unpack_sequence, unpickle, unpickle_list
• 📄table
• 📈time plot

vs. 3.13.0

• Geometric mean: 1.070x slower (HPT: reliability of 100.00%, 1.03x slower at 99th %ile)
• Memory usage: 1.21x
• missing benchmarks: chameleon, dask, djangocms, gevent_hub, gunicorn, tornado_http
• new benchmarks: asyncio_tcp, asyncio_tcp_ssl, pickle, pickle_dict, pickle_list, unpack_sequence, unpickle, unpickle_list
• 📄table
• 📈time plot

vs. base

• Geometric mean: 1.149x slower (HPT: reliability of 100.00%, 1.11x slower at 99th %ile)
• Memory usage: 1.14x
• missing benchmarks: 🔴 dask
• 🧠memory plot
• 📄table
• 📈time plot